CASE REPORT

Gardnerella vaginalis as a Rare Cause of Prosthetic Joint Infection

Gautier Hoarau,a Sylvain Bernard,b Patricia Pavese,b Dominique Saragaglia,c Jacques Croize,a and Max Maurina
Department of Clinical Microbiology,a Department of Infectious Diseases,b and Department of Orthopedic Surgery,c University Joseph Fourier Grenoble 1, Grenoble,
France

We describe a septic loosening of a hip prosthesis in a 71-year-old woman caused by Gardnerella vaginalis. Infection was confirmed by culture and molecular identification of this bacterium. The patient was treated by a one-step exchange of prosthesis
and antibiotic therapy combining trimethoprim-sulfamethoxazole and rifampin, with favorable evolution.

71-year-old woman was hospitalized in the infectious diseases unit of Grenoble University Hospital for prolonged fever, altered general status, and pain of the right hip with cruralgia
for 3 months. Her medical history included tuberculosis, diphtheria, hepatitis A, and dyslipidemia for which she received rosuvastatin. The patient had undergone replacement surgery of both
hips 10 years earlier. A septic loosening of the right hip prosthesis
was evoked.
On admission, the patient was subfebrile at 38C. Pain of the
right hip and cruralgia were not relieved by analgesic treatment, including morphine and nonsteroidal anti-inflammatory drugs (visual analogue score, 6/10). A limitation of hip
joint mobility was noted. Walking was possible but with a limp.
There were no motor or sensory deficits, and tendon reflexes
were normal and symmetrical. The rest of the physical examination was unremarkable. There were no clinical signs suggestive of localized infection, especially no signs of pulmonary or
urinary tract infection.
Laboratory tests showed anemia (hemoglobin, 11 g/dl) and
inflammation (C-reactive protein, 70 mg/liter; ferritinemia,
489 g/liter; increased production of alpha 1 and 2 globulins
on serum protein electrophoresis but immunoglobulins within
normal limits). The leukocyte count was 5.2 109 cells/liter.
Cultures of blood and urine samples were negative. A chest
radiograph was normal. A computed tomography (CT) scan
(chest-abdomen-pelvis) was performed and did not show any
cancer or abscess. No fluid effusion was found around the right
hip prosthesis by ultrasonography, and hip radiographs were
normal. However, fluorine-18 fluorodeoxyglucose (FDG) scintigraphy revealed intense uptake around the hip prosthesis that
was highly suggestive of a septic process.
Debridement and one-step surgical replacement of the right
hip prosthesis were performed. During the intervention, the
surgeon noticed a turbid synovial fluid and confirmed the loosening of the femoral component of the hip prosthesis. A new
prosthesis was reimplanted, the process corresponding to uncemented total hip arthroplasty with a dual-mobility acetabular cup. Fifteen clinical specimens were sampled during the
intervention for microbiological analysis: the femoral cortex (1
specimen), the femoral centromedullary channel (4 specimens), the articular fluid (1 specimen), synovial biopsy specimens (5 specimens), and the acetabulum (4 specimens). Direct
examination of Gram-stained smears prepared from clinical
samples was negative in all cases. The 15 specimens were inoc-

4154

jcm.asm.org

Journal of Clinical Microbiology

ulated onto blood agar plates, including Columbia-5% sheep

blood agar plates (bioMrieux, Marcy lEtoile, France) and chocolate-Polyvitex agar plates (bioMrieux), and incubated at 37C under
a 5% CO2 atmosphere for 7 days. They were also inoculated into
Schaedler broth medium (bioMrieux) incubated 14 days at 37C for
cultivation of fastidious bacteria, especially anaerobes. Bacterial
growth in the Schaedler broth medium was evaluated by Gram staining, and positive broths were subcultured on blood agar plates incubated at 37C, either under a 5% CO2 atmosphere or in an anaerobic
atmosphere. A Gram-positive coryneform bacterium was grown
from all clinical specimens. Because identification of Corynebacterium spp. is often unreliable using conventional techniques, we directly performed 16S rRNA gene amplification and sequencing,
using primers fD1 (5=-AGAGTTTGATCCTGGCTCAG-3=) and
rP2 (5=-ACGGCTACCTTGTTACGACTT-3=). A definite identification of Gardnerella vaginalis was obtained within 1 week, with
100% identity of the 1,187-bp amplified sequence (GenBank accession no. JX391978) with that of a type strain of G. vaginalis
(ATCC 14019, GenBank accession number CP002104.1) using
BLAST software (NCBI, Bethesda, MD).
There is currently no standardized protocol for G. vaginalis
antimicrobial susceptibility testing. Thus, we used an agar disk
diffusion method, previously described for coryneform bacteria
(9) (Mueller-Hinton plate, supplemented with 5% sheep blood,
bioMrieux, Lyon, France), and breakpoints advocated by the
CLSI (Clinical and Laboratory Standards Institute) for Streptococcus spp. (4) The isolated strain was susceptible to penicillin G,
amoxicillin, cephalothin, cefotaxime, imipenem, gentamicin,
chloramphenicol, doxycycline, erythromycin, pristinamycin, rifampin, trimethoprim (TMP)-sulfamethoxazole (SMX), and vancomycin but was resistant to ofloxacin.
The patient first received intravenous piperacillin (4 g/day)tazobactam (500 mg/day) plus vancomycin (2 g/day) therapy
for 10 days. When bacterial identification was available, the
treatment was changed to trimethoprim-sulfamethoxazole
(TMP, 320 mg/day; SMX, 1,600 mg/day) and rifampin (1,200
mg/day) for 3 months of oral therapy. The patient had no clin-

Received 27 July 2012 Returned for modification 24 August 2012

Accepted 9 September 2012
Published ahead of print 19 September 2012
Address correspondence to Max Maurin, mmaurin@chu-grenoble.fr.
Copyright 2012, American Society for Microbiology. All Rights Reserved.
doi:10.1128/JCM.01969-12

p. 4154 4156

December 2012 Volume 50 Number 12

Downloaded from http://jcm.asm.org/ on September 28, 2015 by guest

CASE REPORT

Case Report

TABLE 1 Cases of osteoarticular infections caused by Gardnerella vaginalis reported in the literature
Yr of presentation

Age (yr)

Gendera

Lesion

Reference

1986
1995
2004
2008
2012

Neonate
50
48
38
71

M
F
F
F
F

Osteomyelitis of parietal bone

Vertebral disk space infection
Hip arthritis
Discitis
Hip prosthesis infection

Nightingale (11)
Hodge (6)
Sivadon-Tardy (14)
Graham (5)
Present report

M, male; F, female.

G. vaginalis is a facultative anaerobic bacterium that occasionally colonizes the vaginal mucosa in women (3). It has also
been detected in the microflora of other mucosa, especially of
the oral cavity (15). It is mainly associated with bacterial vaginosis (3). Rarely, G. vaginalis has been associated with bacteremia, especially in the context of postpartum (12) and gynecological (3, 12) infections, and neonatal infections (3).
Bloodstream infections in men, more rare, have been also described (3, 7). Reports on G. vaginalis osteoarticular infections
are exceptional and, not surprisingly, occurred in women or
neonates. Two nosological forms have been described: the first
one is associated with a reactive arthritis (13), and it is characterized by a negative culture of joint fluid; the second one corresponds to a true osteoarticular infection as demonstrated by
the ability to isolate G. vaginalis from clinical samples (especially joint fluid). Only a few cases of G. vaginalis-related septic
osteoarticular infections have been reported in the literature
(Table 1), including one case of osteomyelitis of the parietal
bone in a neonate (11), two cases of discitis in women (5, 6),
and one case of hip arthritis in a woman (14). In all these cases,
no genital infection could be demonstrated.
G. vaginalis infection of the prosthetic hip in our patient
cannot be questioned because the bacterium was isolated from
all 15 clinical samples collected during surgery. No other microorganism was isolated from the same specimens. The pathogenic implication of G. vaginalis in the septic loosening of the
hip is highly likely, according to the OSIRIS (Oxford Skeletal
Infection Research and Intervention Service) collaborative
study group criteria (1). This infection occurred 10 years after
the first surgical replacement and thus is a delayed complication of the initial arthroplasty. As is often the case in very late
infections, the clinical presentation was insidious, with progressive worsening of joint pain and without evidence of severe
sepsis (10). Infection of the hip prosthesis probably occurred
after transient hematogenous spread of G. vaginalis, which
probably explains the fact that blood cultures were sterile at the
time of hospitalization. It should be mentioned, however, that
G. vaginalis growth is partially inhibited by sodium polyanethol sulfonate (SPS), an anticoagulant present in blood culture bottles. The patient had no predisposing factors for infections, such as immunosuppression (no HIV infection, no

December 2012 Volume 50 Number 12

corticosteroid intake), diabetes mellitus, renal failure, obesity,

or alcohol or tobacco consumption. Previous hip arthroplasty
was the only risk factor per se for prosthetic joint infection (10).
Dental care was performed 1 month before the onset of infection and could potentially lead to G. vaginalis bacteremia, since
this bacterium has been previously detected in the oral cavity
(15), but this possibility remains to be established. Berbari et al.
(2) previously reported that dental procedures were not associated with an increased risk of prosthetic hip infection, and
antibiotic prophylaxis in patients with hip arthroplasty and
undergoing dental treatment did not decrease the risk of septic
loosening (2).
The formation of a bacterial biofilm around the prosthesis is
strongly associated with a septic loosening outcome (16). With the
experience of bacterial vaginosis, we know that G. vaginalis can
form a dense biofilm and possesses adhesion factors (3). Interestingly, Marrazzo et al. (8) recently demonstrated that bacterial
vaginosis could be acquired from extravaginal reservoirs. These
data suggest the ability of G. vaginalis to colonize surfaces outside
its vaginal ecological niche.
Our case is the first report of G. vaginalis prosthesis joint infection. It emphasizes that G. vaginalis may cause extravaginal infections, especially osteoarticular infections in patients with or without joint prosthesis. It is important for microbiologists and
physicians to be aware of this possibility, since G. vaginalis is a
fastidious and difficult-to-identify microorganism.
ACKNOWLEDGMENT
There are no conflicts of interest to disclose.

REFERENCES
1. Atkins BL, et al. 1998. Prospective evaluation of criteria for microbiological diagnosis of prosthetic-joint infection at revision arthroplasty. The
OSIRIS Collaborative Study Group. J. Clin. Microbiol. 36:29322939.
2. Berbari EF, et al. 2010. Dental procedures as risk factors for prosthetic hip
or knee infection: a hospital-based prospective case-control study. Clin.
Infect. Dis. 50:8 16.
3. Catlin BW. 1992. Gardnerella vaginalis: characteristics, clinical considerations, and controversies. Clin. Microbiol. Rev. 5:213237.
4. Clinical and Laboratory Standards Institute. 2009. Performance standards for antimicrobial disk susceptibility tests; approved standard
tenth edition. CLSI document M02-A10. Clinical and Laboratory Standards Institute, Wayne, PA.
5. Graham S, Howes C, Dunsmuir R, Sandoe J. 2009. Vertebral osteomyelitis and discitis due to Gardnerella vaginalis. J. Med. Microbiol. 58:1382
1384.
6. Hodge TW, Jr, Levy CS, Smith MA. 1995. Disk space infection due to
Gardnerella vaginalis. Clin. Infect. Dis. 21:443 445.
7. Lagac-Wiens PR, et al. 2008. Gardnerella vaginalis bacteremia in a previously healthy man: case report and characterization of the isolate. J. Clin.
Microbiol. 46:804 806.
8. Marrazzo JM, et al. 2012. Extravaginal reservoirs of vaginal bacteria as
risk factors for incident bacterial vaginosis. J. Infect. Dis. 205:1580 1588.

jcm.asm.org 4155

Downloaded from http://jcm.asm.org/ on September 28, 2015 by guest

ical signs of genital infection. G. vaginalis was not isolated from

vaginal and urine specimens. We did not look for other carriage
sites. Following prosthetic replacement, the evolution has been
satisfactory for the 6-month follow-up, with disappearance of
fever, pain in the hip, and cruralgia and normalization of the
C-reactive protein level.

Case Report

9. Martnez-Martnez L, Ortega MC, Surez AI. 1995. Comparison of

E-test with broth microdilution and disk diffusion for susceptibility testing of coryneform bacteria. J. Clin. Microbiol. 33:1318 1321.
10. Moran E, Byren I, Atkins BL. 2010. The diagnosis and management of prosthetic
joint infections. J. Antimicrob. Chemother. 65(Suppl 3):iii45iii54.
11. Nightingale LM, et al. 1986. Cephalhematoma complicated by osteomyelitis presumed due to Gardnerella vaginalis. JAMA 256:1936 1937.
12. Reimer LG, Reller LB. 1984. Gardnerella vaginalis bacteremia: a review of
thirty cases. Obstet. Gynecol. 64:170 172.

13. Schapira D, Braun-Moscovici Y, Nahir AM. 2002. Reactive arthritis

induced by Gardnerella vaginalis. Clin. Exp. Rheumatol. 20:732733.
14. Sivadon-Tardy V, et al. 2009. Gardnerella vaginalis acute hip arthritis in
a renal transplant recipient. J. Clin. Microbiol. 47:264 265.
15. Zaura E, Keijser BJ, Huse SM, Crielaard W. 2009. Defining the healthy
core microbiome of oral microbial communities. BMC Microbiol.
9:259.
16. Zimmerli W, Trampuz A, Ochsner PE. 2004. Prosthetic-joint infections.
N. Engl. J. Med. 351:16451654.

Downloaded from http://jcm.asm.org/ on September 28, 2015 by guest

4156

jcm.asm.org

Journal of Clinical Microbiology